'Sexed-up' numbers should not always be accepted as science

Times Online:

In December 2005 a study in the journal Nature offered the
observation that the circulation in the North Atlantic Ocean, which sustains
the Gulf Stream, had weakened by up to 30 per cent over the previous few
decades. This figure and its juxtapositioning alongside the melodrama of
films such as The Day after Tomorrow were amplified through the
cooperation of scientists and media to result in headlines such as “Alarm
over dramatic weakening of Gulf Stream” ( The Guardian, Dec
1, 2005). The urban myth that emerged from this episode was that we were
closer to a mini Ice Age in the UK than had previously been thought.
Eighteen months later, however, and unremarked by the media, two studies in
equally reputable journals pointed out that such a trend was within the
range of natural variability and may signify nothing at all.

A second example concerns the claim that, “by the end of this century,
climate change will have killed around 182 million people in sub-Saharan
Africa” (Christian Aid, May 2006). This number – 180 million African dead –
has become one of the most widely cited numbers in the litany of doom that
accompanies talk of climate change. In this case, however, the number 180
million was sexed-up science. Christian Aid took the worst-case climate
scenario, the highest population scenario and the scenario with the least
public health intervention and conjured the number into being. And here it
has stayed, a number detached from its receding scientific origins in which
assumptions were overlain on scenarios that captured uncertainties.

So… according to the headline, we should only *sometimes* accept exaggerated and bogus numbers as scientific fact?